Relationship Intelligence in Meta Glasses
We built relationship intelligence into Meta glasses using the Steer AI MCP at the TechTO Personal Agent Hackathon for Toronto Tech Week. Aleksander Strazisar, Campbell Bird, Anita So, and Tyler Squibb spent all day Sunday putting the demo together during the hackathon, and the result went even better than expected. It was great hearing from so many people at the event that they thought the concept was cool, useful, and worth continuing to build on.
The demo showed how Meta glasses can work with the Steer MCP to surface relationship context in the moment. From Tyler Squibb's perspective, when he looked at someone he knew who had agreed to be included, Steer could help provide context about who they are, how he knows them, and why that relationship matters. That is the part we are excited about: relationship intelligence becoming available at the exact moment it is useful, without needing to search through notes, tabs, profiles, or memory.
Why We Built It
Before the hackathon, Tyler Squibb had heard from many people that they wanted something like this. People kept describing the same problem in different ways: they wanted better context before conversations, faster recall when meeting someone, and a more natural way to understand who is in front of them without interrupting the flow of the interaction. The TechTO Personal Agent Hackathon felt like the right time and place to build a first version.
Steer already helps people and teams understand their relationships, find warm paths, and give AI systems better context through MCP. Bringing that into Meta glasses was a natural extension of the same vision, because relationship intelligence should meet people where the conversation is happening. The hackathon gave us a chance to take that idea out of a normal desktop workflow and test what it could feel like in a more real-world setting.
What Comes Next
This was a hackathon build, but it gave us a strong foundation to keep improving. We plan to continue building on what we created, improve the experience, and explore what relationship-aware agents can look like when they move beyond desktop workflows and into real-world interactions. The response from people at the event was encouraging, and it reinforced that there is real interest in tools that help people be more prepared, more thoughtful, and more context-aware in the moments that matter.
Thank You
Thank you to all the sponsors who helped make the event possible: Rootly, Tangerine, ElevenLabs, Motion (Creative Analytics), Backboard.io, and Codalio. Thank you as well to Adrian Lee Kwen, Bryson Lee-Kwen, and Riddhi Mandal for the help and assistance throughout the hackathon. Reach out if you would like to see more.